* Inside the mind of a great data center manager
The TV show “Monk” features a detective who’s obsessive-compulsive. He gets bent out of shape over a picture hanging off-center on a wall, or papers that aren’t lined up 100% evenly on a desk. The joke is that his attention to detail is what makes him a great detective (and an impossible person to work with).
It also brings up an excellent question: What’s the personality profile that makes for a terrific data center manager? Here are a few characteristics that world-class data center managers typically share (above and beyond technical excellence and engineering wizardry, of course):
* Obsession with consistency and perfection. Ninety-nine or 99.9% reliability is good enough for some parts of an IT organization. If a server has to be rebooted every now and then, or if an application has a sporadic bug, users can learn to live with it. But data center managers assume the responsibility of keeping their infrastructure “always on,” delivering the same level of availability as power and water utilities. Think about it: A service that’s up 99% of the time is actually down almost four days out of the year – longer than the East Coast power blackout in August 2003. Even the vaunted “five nines” of reliability amounts to more than five minutes per year of outages – a delivery level that’s not acceptable for data centers supporting, say, an air traffic control center or a stock exchange. There are runners who haven’t missed a day of running in five, 10, and even 15 years. They run through weddings and funerals, in sickness and in health, in rain, sleet, snow, and sweltering sun. They’d probably make outstanding data center managers.
* Focus on process. There are people who simply aren’t content with solving a problem – they want to find out why the problem occurred, and fix the problem that caused the problem. Nine times out of 10 the real issue is process: technology worked as expected, but humans did something that they shouldn’t have. My favorite example: The AT&T frame relay outage in 1998 occurred because an engineer loaded untested software patches – which were predictably buggy – into production switches. The process fix? Don’t load untested software into production machines. Great data center managers are obsessed with finding and fixing process flaws.
* Visionary brilliance. It may seem odd to throw this characteristic in with the above two. But truly great data center managers are always looking at ways to improve performance and do things that haven’t been done before, particularly when there’s a tangible benefit to be reaped. It’s why large financial firms got into grid computing long before the rest of us – crunching numbers better than the competition can yield a direct impact on the company’s bottom line (not to mention the data center manager’s bonus). Great data center managers seize on promising new technology, test it thoroughly, understand its process requirements – and put it to work.
It’s not often that you find humans with this combination of skills – but then, true greatness is rare in any profession.
